To answer a question  from David which I think is  of interest to everybody:
the minimum storage size for LISTSERV is 2M, and the recommended size is 3M if
you plan to process large files or if you have a lot of lists/users.
 
  Now even  with 4M you  can get  messages saying "out  of storage" and  a CMS
abend. This  is not  LISTSERV's fault. Somewhere  in the code  it calls  a CMS
utility  (GLOBALV?  DMSDDL? EXECIO?  Or,  more  likely, COPYFILE...)  which  I
haven't been  able to  identify and  which issues  a GETMAIN  call. I  find it
unacceptable that  a CMS utility should  use GETMAIN and not  DMSFREE. GETMAIN
reports that there is no storage left because of the fragmentation and because
of  the fact  that it  can only  work  on a  contiguous storage  chunk at  the
beginning of  free storage.  If I  use a lot  of stack  storage, then  issue a
GLOBALV SET and purge the stack storage, I may have drastically diminished the
storage  available  to DMSSMN  (FREEMAIN/GETMAIN)  without  having eaten  much
storage in fact.
 
  The worst is that this is an  unconditional GETMAIN which causes an abend if
not satisfied. I  would really like to  know what causes this. I  am NOT using
any of  the OS simulation crap  (FILEDEF, MOVEFILE, LISTDS, AMSERV),  only CMS
commands. Any idea?
 
  Anyway the point  is that even with  8M, you can still get  these abends. No
use increasing  storage above 3M. Just  reboot LISTSERV and CH  LISTSERV R ALL
NOHOLD to restart the failing file.
 
  Eric